ESC 0/1-hour hs-cTnT algorithm does not safely exclude 30-day cardiac events in known CAD
Performance of the European Society of Cardiology 0/1-Hour Algorithm With High-Sensitivity Cardiac Troponin T Among Patients With Known Coronary Artery DiseaseLiterature - Ashburn NP, Snavely AC, O’Neill JC, et al. - JAMA Cardiol. 2023 Apr 1;8(4):347-356. doi: 10.1001/jamacardio.2023.0031
Introduction and methods
The European Society of Cardiology (ESC) 0/1-hour algorithm is an accelerated diagnostic protocol that is designed to rapidly rule in or rule out NSTEMI based on hs-cTnT level . Previous studies, which were conducted primarily in Europe, showed that with this algorithm, almost 60% of patients are classified as “rule-out”, with a negative predictive value (NPV) of ≥99% [2-8]. However, limited data exist regarding its performance in patients with known coronary artery disease (CAD), particularly in a US cohort of chest pain patients.
Aim of the study
The study objectives were: (1) to determine whether the ESC 0/1-hour hs-cTnT algorithm could achieve the NPV threshold of 99% required to safely rule out 30-day cardiac death or MI among patients with known CAD; and (2) to evaluate and compare the diagnostic performance of this algorithm for 30-day MACE in patients with and with no known CAD.
This was a preplanned subgroup analysis of the STOP-CP (High-Sensitivity Cardiac Troponin T (Gen 5 STAT assay) to Optimize Chest Pain Risk Stratification) study. In this prospective multicenter cohort study, patients aged ≥21 years with acute chest pain or other ACS-related symptoms were enrolled at 8 (mostly academic) emergency departments in the USA. Presence of STEMI was an exclusion criterion. Known CAD—defined as prior MI, coronary revascularization, or coronary stenosis ≥70%—was determined by medical record review and patient self-reporting. In this analysis, 1430 patients were included, of whom 449 (31.4%) had known CAD.
Serial blood samples were collected for hs-cTnT measurement at baseline and 1 hour thereafter. hs-cTnT results were used to stratify patients into 3 zones: rule-out (0-hour hs-cTnT<6 ng/L or 0-hour <12 ng/L and ∆0/1-hour <3 ng/L), observation (any hs-cTnT value or ∆ not meeting rule-out or rule-in criteria), and rule-in (0-hour hs-cTnT ≥52 ng/L or ∆0/1-hour ≥5 ng/L). Treating clinicians were blinded to the hs-cTnT results, and patient care was guided by contemporary cTn results.
The primary endpoint was cardiac death or MI at 30 days. Secondary endpoints included MACE (defined as cardiac death, MI, or coronary revascularization) at 30 days; individual MACE subcomponents; and efficacy (defined as proportion of patients classified into rule-out zone).
- Usage of the ESC 0/1-hour algorithm classified 178 of the 449 patients with known CAD (39.6%) into the rule-out zone compared with 648 of the 981 patients with no known CAD (66.1%) (P<0.001).
- By day 30, cardiac death or MI had occurred in 6 of 178 rule-out patients with CAD (3.4%) and 7 of 648 rule-out patients with no CAD (1.1%) (OR: 3.19; 95%CI: 1.05–9.63; P=0.04).
- The NPV for 30-day cardiac death or MI was 96.6% (95%CI: 92.8%–98.8%) in patients with CAD and 98.9% (95%CI: 97.8%–99.6%) in patients with no CAD (P=0.04). The sensitivity for 30-day cardiac death or MI was similar in both groups (93.2%; 95%CI: 85.7%–97.5% in the group with CAD vs. 92.6%; 95%CI: 85.4%–97.0% in the group without CAD; P=1.0).
- MACE at 30 days occurred in 14 of 178 rule-out patients with CAD (7.9%) vs. 9 of 648 rule-out patients with no CAD (1.4%) (OR: 6.06; 95%CI: 2.59–14.25; P<0.001).
- The NPV for 30-day MACE was 92.1% (95%CI: 87.1%–95.6%) in patients with CAD and 98.6% (95%CI: 97.4%–99.4%) in patients with no CAD. The sensitivity for 30-day MACE was 86.5% (95%CI: 78.4%–92.4%) and 90.9% (95%CI: 83.4%–95.8%), respectively.
- Usage of the ESC 0/1-hour algorithm classified 88 of the 449 patients with CAD (19.6%) into the rule-in zone compared with 102 of the 981 patients with no CAD (10.4%) (P<0.001).
- The incidence of cardiac death or MI at 30 days was similar for rule-in patients with and with no CAD (52/88 (59.1%) vs. 59/102 (57.8%); OR: 1.05; 95%CI: 0.59–1.88; P=0.88).
- The positive predictive value (PPV) for 30-day cardiac death or MI was 59.1% (95%CI: 48.1%–69.4%) in patients with CAD and 57.8% (95%CI: 47.7%–67.6%) in patients with no CAD (P=0.88). The positive likelihood ratio was lower for patients with CAD (5.9; 95%CI: 4.2–8.5) compared with those with no CAD (12.8; 95%CI: 9.2–17.8; P=0.002).
- As there were no revascularizations in rule-in patients, the OR and PPV for 30-day MACE were the same as those for 30-day cardiac death or MI.
Interaction between algorithm classification and known CAD
- Multivariable logistic regression analysis indicated there was an interaction between the ESC 0/1-hour algorithm and known CAD for 30-day MACE (P=0.006) but not for 30-day cardiac death or MI (P=0.20).
- Known CAD was associated with an increased rate of 30-day MACE (compared with no known CAD) in patients classified to the rule-out zone with the ESC 0/1-hour algorithm (adjusted OR: 6.08; 95%CI: 2.51–14.72). However, 30-day MACE rates were similar in patients with or with no known CAD for those classified to the observation zone (adjusted OR: 1.71; 95%CI: 0.99–2.95) or rule-in zone (adjusted OR: 1.07; 95%CI: 0.54–2.09).
In this preplanned subgroup analysis of the observational STOP-CP study, the ESC 0/1-hour hs-cTnT algorithm had an NPV below the accepted 99% threshold for 30-day cardiac death or MI in patients presenting at a US emergency department with acute chest pain or other ACS-related symptoms who had a history of CAD. Although the NPV was significantly higher in patients with no known CAD, it still did not meet the threshold. An NPV <99% was also observed for 30-day MACE in both patient groups. According to the authors, their results suggested the ESC 0/1-hour hs-cTnT algorithm “is likely not safe for routine use among US patients with chest pain [in the emergency department], regardless of CAD status.”